Description of problem: In order for DMA remapping with IOMMU to be supported, a hint regarding driver support should be added to the OS in the INF file: https://docs.microsoft.com/en-us/windows-hardware/drivers/pci/enabling-dma-remapping-for-device-drivers Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info: How to test: * Check https://docs.microsoft.com/en-us/windows-hardware/drivers/pci/enabling-dma-remapping-for-device-drivers device manager screenshot for the DAM remapping capability of the installed device. * Applicable for Windows 10 (Windows Server 2016) and up
https://github.com/virtio-win/kvm-guest-drivers-windows/pull/647
I test with 211 driver with "iommu device + virtio netkvm device" ,found the guest can not get network , and qmo pop up error message: (qemu) qemu-kvm: unable to start vhost net: 12: falling back on userspace virtio qemu-kvm: virtio: bogus descriptor or out of resources qemu command line: /usr/libexec/qemu-kvm \ -name 'avocado-vt-vm3' \ -machine q35,kernel-irqchip=split \ -nodefaults \ -vga std \ -device pcie-root-port,port=0x10,chassis=1,id=pci.1,bus=pcie.0,multifunction=on,addr=0x3 \ -device pcie-root-port,port=0x11,chassis=2,id=pci.2,bus=pcie.0,addr=0x3.0x1 \ -device pcie-root-port,port=0x12,chassis=3,id=pci.3,bus=pcie.0,addr=0x3.0x2 \ -device pcie-root-port,port=0x13,chassis=4,id=pci.4,bus=pcie.0,addr=0x3.0x3 \ -device pcie-root-port,port=0x14,chassis=5,id=pci.5,bus=pcie.0,addr=0x3.0x4 \ -device pcie-root-port,port=0x15,chassis=6,id=pci.6,bus=pcie.0,addr=0x3.0x5 \ -device pcie-root-port,port=0x16,chassis=7,id=pci.7,bus=pcie.0,addr=0x3.0x6 \ -device pcie-root-port,port=0x17,chassis=8,id=pci.8,bus=pcie.0,addr=0x3.0x7 \ -device virtio-scsi-pci,id=virtio_scsi_pci1,bus=pci.2 \ -blockdev driver=file,cache.direct=off,cache.no-flush=on,filename=/home/test/win2022-64-virtio-scsi.qcow2,node-name=data_node \ -blockdev driver=qcow2,node-name=data_disk,file=data_node \ -device scsi-hd,drive=data_disk,id=disk1,bus=virtio_scsi_pci1.0,serial=kk \ -m 14336 \ -smp 2,maxcpus=4 \ -cpu 'Skylake-Server',hv_stimer,hv_synic,hv_vpindex,hv_relaxed,hv_spinlocks=0x1fff,hv_vapic,hv_time,hv_frequencies,hv_runtime,hv_tlbflush,hv_reenlightenment,hv_stimer_direct,hv_ipi,+kvm_pv_unhalt \ -device piix3-usb-uhci,id=usb -device usb-tablet,id=input0 \ -vnc :10 \ -rtc base=localtime,clock=host,driftfix=slew \ -boot order=cdn,once=c,menu=off,strict=off \ -enable-kvm \ -qmp tcp:0:1231,server,nowait \ -monitor stdio \ -drive file=/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd,if=pflash,format=raw,unit=0,readonly=on \ -drive file=/home/OVMF_VARS.fd,if=pflash,format=raw,unit=1 \ -object rng-builtin,id=builtin-f4hKqkqk \ -device virtio-rng-pci,id=virtio-rng-pci-LilC5UtH,rng=builtin-f4hKqkqk,bus=pci.8,iommu_platform=on,ats=on \ -device intel-iommu,intremap=on,eim=on,device-iotlb=on \ -device virtio-net-pci,mac=9a:36:83:b6:3d:05,id=idJVpmsF,netdev=id23ZUK6,bus=pci.3 \ -netdev tap,id=id23ZUK6,vhost=on \ Addition info: 1. remove -device intel-iommu,intremap=on,eim=on,device-iotlb=on \ , guest can get network 2. just change -device virtio-net-pci to -device e1000e , not hit the issue described above 3. test with 210 driver, not hit this issue, so it should be a regression issue caused by this patch. Thanks Menghuan
(In reply to menli from comment #3) > I test with 211 driver with "iommu device + virtio netkvm device" ,found > the guest can not get network , > > and qmo pop up error message: > (qemu) qemu-kvm: unable to start vhost net: 12: falling back on userspace > virtio > qemu-kvm: virtio: bogus descriptor or out of resources > > > qemu command line: > /usr/libexec/qemu-kvm \ > -name 'avocado-vt-vm3' \ > -machine q35,kernel-irqchip=split \ > -nodefaults \ > -vga std \ > -device > pcie-root-port,port=0x10,chassis=1,id=pci.1,bus=pcie.0,multifunction=on, > addr=0x3 \ > -device > pcie-root-port,port=0x11,chassis=2,id=pci.2,bus=pcie.0,addr=0x3.0x1 \ > -device > pcie-root-port,port=0x12,chassis=3,id=pci.3,bus=pcie.0,addr=0x3.0x2 \ > -device > pcie-root-port,port=0x13,chassis=4,id=pci.4,bus=pcie.0,addr=0x3.0x3 \ > -device > pcie-root-port,port=0x14,chassis=5,id=pci.5,bus=pcie.0,addr=0x3.0x4 \ > -device > pcie-root-port,port=0x15,chassis=6,id=pci.6,bus=pcie.0,addr=0x3.0x5 \ > -device > pcie-root-port,port=0x16,chassis=7,id=pci.7,bus=pcie.0,addr=0x3.0x6 \ > -device > pcie-root-port,port=0x17,chassis=8,id=pci.8,bus=pcie.0,addr=0x3.0x7 \ > -device virtio-scsi-pci,id=virtio_scsi_pci1,bus=pci.2 \ > -blockdev > driver=file,cache.direct=off,cache.no-flush=on,filename=/home/test/win2022- > 64-virtio-scsi.qcow2,node-name=data_node \ > -blockdev driver=qcow2,node-name=data_disk,file=data_node \ > -device > scsi-hd,drive=data_disk,id=disk1,bus=virtio_scsi_pci1.0,serial=kk \ > -m 14336 \ > -smp 2,maxcpus=4 \ > -cpu > 'Skylake-Server',hv_stimer,hv_synic,hv_vpindex,hv_relaxed, > hv_spinlocks=0x1fff,hv_vapic,hv_time,hv_frequencies,hv_runtime,hv_tlbflush, > hv_reenlightenment,hv_stimer_direct,hv_ipi,+kvm_pv_unhalt \ > -device piix3-usb-uhci,id=usb -device usb-tablet,id=input0 \ > -vnc :10 \ > -rtc base=localtime,clock=host,driftfix=slew \ > -boot order=cdn,once=c,menu=off,strict=off \ > -enable-kvm \ > -qmp tcp:0:1231,server,nowait \ > -monitor stdio \ > -drive > file=/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd,if=pflash,format=raw,unit=0, > readonly=on \ > -drive file=/home/OVMF_VARS.fd,if=pflash,format=raw,unit=1 \ > -object rng-builtin,id=builtin-f4hKqkqk \ > -device > virtio-rng-pci,id=virtio-rng-pci-LilC5UtH,rng=builtin-f4hKqkqk,bus=pci.8, > iommu_platform=on,ats=on \ > -device intel-iommu,intremap=on,eim=on,device-iotlb=on \ > -device > virtio-net-pci,mac=9a:36:83:b6:3d:05,id=idJVpmsF,netdev=id23ZUK6,bus=pci.3 \ > -netdev tap,id=id23ZUK6,vhost=on \ > > Addition info: > 1. remove -device intel-iommu,intremap=on,eim=on,device-iotlb=on \ , guest > can get network > 2. just change -device virtio-net-pci to -device e1000e , not hit the issue > described above > 3. test with 210 driver, not hit this issue, so it should be a regression > issue caused by this patch. > > > Thanks > Menghuan Hello Menghuan, Why virtio-net device done't have ",iommu_platform=on,ats=on" in the command line? Best regards, Yan.
(In reply to Yvugenfi from comment #4) > (In reply to menli from comment #3) > > I test with 211 driver with "iommu device + virtio netkvm device" ,found > > the guest can not get network , > > > > and qmo pop up error message: > > (qemu) qemu-kvm: unable to start vhost net: 12: falling back on userspace > > virtio > > qemu-kvm: virtio: bogus descriptor or out of resources > > > > > > qemu command line: > > /usr/libexec/qemu-kvm \ > > -name 'avocado-vt-vm3' \ > > -machine q35,kernel-irqchip=split \ > > -nodefaults \ > > -vga std \ > > -device > > pcie-root-port,port=0x10,chassis=1,id=pci.1,bus=pcie.0,multifunction=on, > > addr=0x3 \ > > -device > > pcie-root-port,port=0x11,chassis=2,id=pci.2,bus=pcie.0,addr=0x3.0x1 \ > > -device > > pcie-root-port,port=0x12,chassis=3,id=pci.3,bus=pcie.0,addr=0x3.0x2 \ > > -device > > pcie-root-port,port=0x13,chassis=4,id=pci.4,bus=pcie.0,addr=0x3.0x3 \ > > -device > > pcie-root-port,port=0x14,chassis=5,id=pci.5,bus=pcie.0,addr=0x3.0x4 \ > > -device > > pcie-root-port,port=0x15,chassis=6,id=pci.6,bus=pcie.0,addr=0x3.0x5 \ > > -device > > pcie-root-port,port=0x16,chassis=7,id=pci.7,bus=pcie.0,addr=0x3.0x6 \ > > -device > > pcie-root-port,port=0x17,chassis=8,id=pci.8,bus=pcie.0,addr=0x3.0x7 \ > > -device virtio-scsi-pci,id=virtio_scsi_pci1,bus=pci.2 \ > > -blockdev > > driver=file,cache.direct=off,cache.no-flush=on,filename=/home/test/win2022- > > 64-virtio-scsi.qcow2,node-name=data_node \ > > -blockdev driver=qcow2,node-name=data_disk,file=data_node \ > > -device > > scsi-hd,drive=data_disk,id=disk1,bus=virtio_scsi_pci1.0,serial=kk \ > > -m 14336 \ > > -smp 2,maxcpus=4 \ > > -cpu > > 'Skylake-Server',hv_stimer,hv_synic,hv_vpindex,hv_relaxed, > > hv_spinlocks=0x1fff,hv_vapic,hv_time,hv_frequencies,hv_runtime,hv_tlbflush, > > hv_reenlightenment,hv_stimer_direct,hv_ipi,+kvm_pv_unhalt \ > > -device piix3-usb-uhci,id=usb -device usb-tablet,id=input0 \ > > -vnc :10 \ > > -rtc base=localtime,clock=host,driftfix=slew \ > > -boot order=cdn,once=c,menu=off,strict=off \ > > -enable-kvm \ > > -qmp tcp:0:1231,server,nowait \ > > -monitor stdio \ > > -drive > > file=/usr/share/edk2/ovmf/OVMF_CODE.secboot.fd,if=pflash,format=raw,unit=0, > > readonly=on \ > > -drive file=/home/OVMF_VARS.fd,if=pflash,format=raw,unit=1 \ > > -object rng-builtin,id=builtin-f4hKqkqk \ > > -device > > virtio-rng-pci,id=virtio-rng-pci-LilC5UtH,rng=builtin-f4hKqkqk,bus=pci.8, > > iommu_platform=on,ats=on \ > > -device intel-iommu,intremap=on,eim=on,device-iotlb=on \ > > -device > > virtio-net-pci,mac=9a:36:83:b6:3d:05,id=idJVpmsF,netdev=id23ZUK6,bus=pci.3 \ > > -netdev tap,id=id23ZUK6,vhost=on \ > > > > Addition info: > > 1. remove -device intel-iommu,intremap=on,eim=on,device-iotlb=on \ , guest > > can get network > > 2. just change -device virtio-net-pci to -device e1000e , not hit the issue > > described above > > 3. test with 210 driver, not hit this issue, so it should be a regression > > issue caused by this patch. > > > > > > Thanks > > Menghuan > > Hello Menghuan, > > Why virtio-net device done't have ",iommu_platform=on,ats=on" in the command > line? > > Best regards, > Yan. Hi Yan， Actually， when I run virtio-rng iommo related case in auto hit error and try to debug it and guess it may related to this patch (I missed iommu_platform=on,ats=on with virtio-net device when I debug),so its a coincidence found the issue described above . I also try 210 driver with the same command line as comment3 , no qemu pop up error message. (qemu) qemu-kvm: unable to start vhost net: 12: falling back on userspace virtio qemu-kvm: virtio: bogus descriptor or out of resources
In 210 build the driver does not claim the compatibility with DMA remapping (iommu), so there problem does not happen. The proper way to run the VM with iommu is to add "iommu_platform=on,ats=on" to every virtio device. Currently only netkvm INF file is instrumented for iommu (in build 211), but more drivers will be instumented in future.
(In reply to ybendito from comment #6) > In 210 build the driver does not claim the compatibility with DMA remapping > (iommu), so there problem does not happen. > The proper way to run the VM with iommu is to add "iommu_platform=on,ats=on" > to every virtio device. > Currently only netkvm INF file is instrumented for iommu (in build 211), but > more drivers will be instumented in future. okay, got it, thanks for clarify it~
Hi Yan, Could you take a look at comment2? Thanks in advance!
(In reply to leidwang from comment #8) > Hi Yan, > > Could you take a look at comment2? Thanks in advance! Yes, you can consider the bug verified.
Hi Yan, Please help to check comment 9，the issue seems to related to this patch , thanks
(In reply to menli from comment #11) > Hi Yan, > > Please help to check comment 9，the issue seems to related to this patch , > thanks The details in comment 9 , only use # nc -C $guest_ip 10022 method to send command can hit the issue. Whether we can verify the bz first and use a separate bz to track it? or other suggestion?
(In reply to menli from comment #12) > (In reply to menli from comment #11) > > Hi Yan, > > > > Please help to check comment 9，the issue seems to related to this patch , > > thanks > > > The details in comment 9 , only use # nc -C $guest_ip 10022 method to send > command can hit the issue. > > Whether we can verify the bz first and use a separate bz to track it? or > other suggestion? Please open another BZ for comment 9. This changes and BZ was for NetKVM only.
(In reply to Yvugenfi from comment #13) > (In reply to menli from comment #12) > > (In reply to menli from comment #11) > > > Hi Yan, > > > > > > Please help to check comment 9，the issue seems to related to this patch , > > > thanks > > > > > > The details in comment 9 , only use # nc -C $guest_ip 10022 method to send > > command can hit the issue. > > > > Whether we can verify the bz first and use a separate bz to track it? or > > other suggestion? > > Please open another BZ for comment 9. > > This changes and BZ was for NetKVM only. actually remove rng device can also detect this issue, So its a regression issue related to "iommu device + virtio netkvm device". update the following details to make it more clear: 1. start guest with following qemu command line: /usr/libexec/qemu-kvm \ -name 'avocado-vt-vm3' \ -machine q35,kernel-irqchip=split \ -nodefaults \ -vga std \ -device pcie-root-port,port=0x10,chassis=1,id=pci.1,bus=pcie.0,multifunction=on,addr=0x3 \ -device pcie-root-port,port=0x11,chassis=2,id=pci.2,bus=pcie.0,addr=0x3.0x1 \ -device pcie-root-port,port=0x12,chassis=3,id=pci.3,bus=pcie.0,addr=0x3.0x2 \ -device pcie-root-port,port=0x13,chassis=4,id=pci.4,bus=pcie.0,addr=0x3.0x3 \ -device pcie-root-port,port=0x14,chassis=5,id=pci.5,bus=pcie.0,addr=0x3.0x4 \ -device pcie-root-port,port=0x15,chassis=6,id=pci.6,bus=pcie.0,addr=0x3.0x5 \ -device pcie-root-port,port=0x16,chassis=7,id=pci.7,bus=pcie.0,addr=0x3.0x6 \ -device pcie-root-port,port=0x17,chassis=8,id=pci.8,bus=pcie.0,addr=0x3.0x7 \ -device virtio-scsi-pci,id=virtio_scsi_pci1,bus=pci.2,disable-legacy=on,disable-modern=false,iommu_platform=on,ats=on \ -blockdev driver=file,cache.direct=off,cache.no-flush=on,filename=/home/win2022-64-virtio-scsi.qcow2,node-name=data_node \ -blockdev driver=qcow2,node-name=data_disk,file=data_node \ -device scsi-hd,drive=data_disk,id=disk1,bus=virtio_scsi_pci1.0,serial=kk \ -m 14336 \ -smp 2,maxcpus=4 \ -cpu 'Skylake-Server',hv_stimer,hv_synic,hv_vpindex,hv_relaxed,hv_spinlocks=0x1fff,hv_vapic,hv_time,hv_frequencies,hv_runtime,hv_tlbflush,hv_reenlightenment,hv_stimer_direct,hv_ipi,+kvm_pv_unhalt \ -device piix3-usb-uhci,id=usb -device usb-tablet,id=input0 \ -vnc :10 \ -rtc base=localtime,clock=host,driftfix=slew \ -boot order=cdn,once=c,menu=off,strict=off \ -enable-kvm \ -qmp tcp:0:1231,server,nowait \ -monitor stdio \ -device intel-iommu,intremap=on,eim=on,device-iotlb=on \ -device virtio-net-pci,mac=9a:36:83:b6:3d:05,id=idJVpmsF,netdev=id23ZUK6,bus=pci.3,disable-legacy=on,disable-modern=false,iommu_platform=on,ats=on \ -netdev tap,id=id23ZUK6,vhost=on \ -drive id=drive_cd1,if=none,snapshot=off,aio=threads,cache=none,media=cdrom,file=/home/kvm_autotest_root/iso/windows/winutils.iso\ -device ide-cd,id=cd2,drive=drive_cd1,bus=ide.0,unit=0 \ 2. connect to guest session: # nc -C $guest_ip 10022 Acutual result: after step2, Return garbled eg: [root@dell-per730-47 ~]# nc -C 10.73.74.208 10022 Please wait... `���Y�a���J0Afd�M����F���� �?���� C:\>ipconfig Windows IP Configuration /�����������'��������������M������E C:\> Additional info: 1.try the same steps above with 210 driver, no return garbled 2. just remove -device intel-iommu,intremap=on,eim=on,device-iotlb=on , can not hit this issue, the result like following, So it should be a regression issue related to "iommu device + virtio netkvm device". [root@dell-per730-47 ~]# nc -C 10.73.74.208 10022 Please wait... Microsoft Windows [Version 10.0.20348.169] (c) Microsoft Corporation. All rights reserved. C:\>ipconfig Windows IP Configuration Ethernet adapter Ethernet Instance 0 2: Connection-specific DNS Suffix . : lab.eng.pek2.redhat.com Link-local IPv6 Address . . . . . : fe80::20a4:b429:19d4:59aa%8 IPv4 Address. . . . . . . . . . . : 10.73.74.208 Subnet Mask . . . . . . . . . . . : 255.255.252.0 Default Gateway . . . . . . . . . : 10.73.75.254
Hi Yan, Would you please help check comment 14 to confirm should we reassign this bz or file a new bz for it? Regards, Qianqian
(In reply to Qianqian Zhu from comment #15) > Hi Yan, > > Would you please help check comment 14 to confirm should we reassign this bz > or file a new bz for it? > > Regards, > Qianqian Please re-assign. Thanks.
(In reply to menli from comment #14) > (In reply to Yvugenfi from comment #13) > > (In reply to menli from comment #12) > > > (In reply to menli from comment #11) > > > > Hi Yan, > > > > > > > > Please help to check comment 9，the issue seems to related to this patch , > > > > thanks > > > > > > > > > The details in comment 9 , only use # nc -C $guest_ip 10022 method to send > > > command can hit the issue. > > > > > > Whether we can verify the bz first and use a separate bz to track it? or > > > other suggestion? > > > > Please open another BZ for comment 9. > > > > This changes and BZ was for NetKVM only. > > actually remove rng device can also detect this issue, So its a regression > issue related to "iommu device + virtio netkvm device". update the following > details to make it more clear: > > > 1. start guest with following qemu command line: > > /usr/libexec/qemu-kvm \ > -name 'avocado-vt-vm3' \ > -machine q35,kernel-irqchip=split \ > -nodefaults \ > -vga std \ > -device > pcie-root-port,port=0x10,chassis=1,id=pci.1,bus=pcie.0,multifunction=on, > addr=0x3 \ > -device > pcie-root-port,port=0x11,chassis=2,id=pci.2,bus=pcie.0,addr=0x3.0x1 \ > -device > pcie-root-port,port=0x12,chassis=3,id=pci.3,bus=pcie.0,addr=0x3.0x2 \ > -device > pcie-root-port,port=0x13,chassis=4,id=pci.4,bus=pcie.0,addr=0x3.0x3 \ > -device > pcie-root-port,port=0x14,chassis=5,id=pci.5,bus=pcie.0,addr=0x3.0x4 \ > -device > pcie-root-port,port=0x15,chassis=6,id=pci.6,bus=pcie.0,addr=0x3.0x5 \ > -device > pcie-root-port,port=0x16,chassis=7,id=pci.7,bus=pcie.0,addr=0x3.0x6 \ > -device > pcie-root-port,port=0x17,chassis=8,id=pci.8,bus=pcie.0,addr=0x3.0x7 \ > -device > virtio-scsi-pci,id=virtio_scsi_pci1,bus=pci.2,disable-legacy=on,disable- > modern=false,iommu_platform=on,ats=on \ > -blockdev > driver=file,cache.direct=off,cache.no-flush=on,filename=/home/win2022-64- > virtio-scsi.qcow2,node-name=data_node \ > -blockdev driver=qcow2,node-name=data_disk,file=data_node \ > -device > scsi-hd,drive=data_disk,id=disk1,bus=virtio_scsi_pci1.0,serial=kk \ > -m 14336 \ > -smp 2,maxcpus=4 \ > -cpu > 'Skylake-Server',hv_stimer,hv_synic,hv_vpindex,hv_relaxed, > hv_spinlocks=0x1fff,hv_vapic,hv_time,hv_frequencies,hv_runtime,hv_tlbflush, > hv_reenlightenment,hv_stimer_direct,hv_ipi,+kvm_pv_unhalt \ > -device piix3-usb-uhci,id=usb -device usb-tablet,id=input0 \ > -vnc :10 \ > -rtc base=localtime,clock=host,driftfix=slew \ > -boot order=cdn,once=c,menu=off,strict=off \ > -enable-kvm \ > -qmp tcp:0:1231,server,nowait \ > -monitor stdio \ > -device intel-iommu,intremap=on,eim=on,device-iotlb=on \ > -device > virtio-net-pci,mac=9a:36:83:b6:3d:05,id=idJVpmsF,netdev=id23ZUK6,bus=pci.3, > disable-legacy=on,disable-modern=false,iommu_platform=on,ats=on \ > -netdev tap,id=id23ZUK6,vhost=on \ > -drive > id=drive_cd1,if=none,snapshot=off,aio=threads,cache=none,media=cdrom,file=/ > home/kvm_autotest_root/iso/windows/winutils.iso\ > -device ide-cd,id=cd2,drive=drive_cd1,bus=ide.0,unit=0 \ > > > 2. connect to guest session: > # nc -C $guest_ip 10022 > > > Acutual result: > > after step2, Return garbled > eg: > > [root@dell-per730-47 ~]# nc -C 10.73.74.208 10022 > Please wait... > `���Y�a���J0Afd�M����F���� �?���� > > > > C:\>ipconfig > > Windows IP Configuration > > /�����������'��������������M������E > C:\> > > > > Additional info: > > 1.try the same steps above with 210 driver, no return garbled > 2. just remove -device intel-iommu,intremap=on,eim=on,device-iotlb=on , can > not hit this issue, the result like following, So it should be a regression > issue related to "iommu device + virtio netkvm device". > > [root@dell-per730-47 ~]# nc -C 10.73.74.208 10022 > Please wait... > Microsoft Windows [Version 10.0.20348.169] > (c) Microsoft Corporation. All rights reserved. > > C:\>ipconfig > > Windows IP Configuration > > > Ethernet adapter Ethernet Instance 0 2: > > Connection-specific DNS Suffix . : lab.eng.pek2.redhat.com > Link-local IPv6 Address . . . . . : fe80::20a4:b429:19d4:59aa%8 > IPv4 Address. . . . . . . . . . . : 10.73.74.208 > Subnet Mask . . . . . . . . . . . : 255.255.252.0 > Default Gateway . . . . . . . . . : 10.73.75.254 Can you please provide the service\application on the guest that opening 10022 port?
hi leidong, Could you please help to check comment 17?
(In reply to Yvugenfi from comment #17) > (In reply to menli from comment #14) > > (In reply to Yvugenfi from comment #13) > > > (In reply to menli from comment #12) > > > > (In reply to menli from comment #11) > > > > > Hi Yan, > > > > > > > > > > Please help to check comment 9，the issue seems to related to this patch , > > > > > thanks > > > > > > > > > > > > The details in comment 9 , only use # nc -C $guest_ip 10022 method to send > > > > command can hit the issue. > > > > > > > > Whether we can verify the bz first and use a separate bz to track it? or > > > > other suggestion? > > > > > > Please open another BZ for comment 9. > > > > > > This changes and BZ was for NetKVM only. > > > > actually remove rng device can also detect this issue, So its a regression > > issue related to "iommu device + virtio netkvm device". update the following > > details to make it more clear: > > > > > > 1. start guest with following qemu command line: > > > > /usr/libexec/qemu-kvm \ > > -name 'avocado-vt-vm3' \ > > -machine q35,kernel-irqchip=split \ > > -nodefaults \ > > -vga std \ > > -device > > pcie-root-port,port=0x10,chassis=1,id=pci.1,bus=pcie.0,multifunction=on, > > addr=0x3 \ > > -device > > pcie-root-port,port=0x11,chassis=2,id=pci.2,bus=pcie.0,addr=0x3.0x1 \ > > -device > > pcie-root-port,port=0x12,chassis=3,id=pci.3,bus=pcie.0,addr=0x3.0x2 \ > > -device > > pcie-root-port,port=0x13,chassis=4,id=pci.4,bus=pcie.0,addr=0x3.0x3 \ > > -device > > pcie-root-port,port=0x14,chassis=5,id=pci.5,bus=pcie.0,addr=0x3.0x4 \ > > -device > > pcie-root-port,port=0x15,chassis=6,id=pci.6,bus=pcie.0,addr=0x3.0x5 \ > > -device > > pcie-root-port,port=0x16,chassis=7,id=pci.7,bus=pcie.0,addr=0x3.0x6 \ > > -device > > pcie-root-port,port=0x17,chassis=8,id=pci.8,bus=pcie.0,addr=0x3.0x7 \ > > -device > > virtio-scsi-pci,id=virtio_scsi_pci1,bus=pci.2,disable-legacy=on,disable- > > modern=false,iommu_platform=on,ats=on \ > > -blockdev > > driver=file,cache.direct=off,cache.no-flush=on,filename=/home/win2022-64- > > virtio-scsi.qcow2,node-name=data_node \ > > -blockdev driver=qcow2,node-name=data_disk,file=data_node \ > > -device > > scsi-hd,drive=data_disk,id=disk1,bus=virtio_scsi_pci1.0,serial=kk \ > > -m 14336 \ > > -smp 2,maxcpus=4 \ > > -cpu > > 'Skylake-Server',hv_stimer,hv_synic,hv_vpindex,hv_relaxed, > > hv_spinlocks=0x1fff,hv_vapic,hv_time,hv_frequencies,hv_runtime,hv_tlbflush, > > hv_reenlightenment,hv_stimer_direct,hv_ipi,+kvm_pv_unhalt \ > > -device piix3-usb-uhci,id=usb -device usb-tablet,id=input0 \ > > -vnc :10 \ > > -rtc base=localtime,clock=host,driftfix=slew \ > > -boot order=cdn,once=c,menu=off,strict=off \ > > -enable-kvm \ > > -qmp tcp:0:1231,server,nowait \ > > -monitor stdio \ > > -device intel-iommu,intremap=on,eim=on,device-iotlb=on \ > > -device > > virtio-net-pci,mac=9a:36:83:b6:3d:05,id=idJVpmsF,netdev=id23ZUK6,bus=pci.3, > > disable-legacy=on,disable-modern=false,iommu_platform=on,ats=on \ > > -netdev tap,id=id23ZUK6,vhost=on \ > > -drive > > id=drive_cd1,if=none,snapshot=off,aio=threads,cache=none,media=cdrom,file=/ > > home/kvm_autotest_root/iso/windows/winutils.iso\ > > -device ide-cd,id=cd2,drive=drive_cd1,bus=ide.0,unit=0 \ > > > > > > 2. connect to guest session: > > # nc -C $guest_ip 10022 > > > > > > Acutual result: > > > > after step2, Return garbled > > eg: > > > > [root@dell-per730-47 ~]# nc -C 10.73.74.208 10022 > > Please wait... > > `���Y�a���J0Afd�M����F���� �?���� > > > > > > > > C:\>ipconfig > > > > Windows IP Configuration > > > > /�����������'��������������M������E > > C:\> > > > > > > > > Additional info: > > > > 1.try the same steps above with 210 driver, no return garbled > > 2. just remove -device intel-iommu,intremap=on,eim=on,device-iotlb=on , can > > not hit this issue, the result like following, So it should be a regression > > issue related to "iommu device + virtio netkvm device". > > > > [root@dell-per730-47 ~]# nc -C 10.73.74.208 10022 > > Please wait... > > Microsoft Windows [Version 10.0.20348.169] > > (c) Microsoft Corporation. All rights reserved. > > > > C:\>ipconfig > > > > Windows IP Configuration > > > > > > Ethernet adapter Ethernet Instance 0 2: > > > > Connection-specific DNS Suffix . : lab.eng.pek2.redhat.com > > Link-local IPv6 Address . . . . . : fe80::20a4:b429:19d4:59aa%8 > > IPv4 Address. . . . . . . . . . . : 10.73.74.208 > > Subnet Mask . . . . . . . . . . . : 255.255.252.0 > > Default Gateway . . . . . . . . . : 10.73.75.254 > > Can you please provide the service\application on the guest that opening > 10022 port? Checked the port 10022 and the corresponding pid, Port 10022 used by rss.exe The screenshot as follows: https://docs.google.com/document/d/1CAvnSNDKMP0hg7W9SW-F0vgrSZXeIZkUDNxLKNqOCBw/edit?usp=sharing
Please set regression flag to this BZ. The change brakes NetKVM with IOMMU by introducing random corruption to the packets.
Hi Yan, Could you adjust the DTM? Since the DTM is too later, our test time will be very tight. We'd better change the DTM to 23 or earlier. Thanks!
Tested netkvm with build 214,there is no regression.Thanks!